Applying least squares principles to estimating sources of contamination in the Neuse River basin

Abstract:

Motivated by applications to water quality monitoring using fluorescence spectroscopy, this talk will develop a source apportionment model for high dimensional profiles of dissolved organic matter (DOM). I will propose simple methods to estimate the parameters of a linear source apportionment model, and I will demonstrate how these estimates relate to the methods of ordinary and generalized least squares. I'll also propose corresponding predictors for missing elements of a DOM profile using this least squares framework. The practical implications of these results for variance reduction will be illustrated using fluorescence spectroscopy data collected from the Neuse River basin in North Carolina.
 

 

Presented By: Dr. Jordan Bryan (UNC)